At an abstract level, people generally use tools and technologies primarily for three reasons, to make money, to save money, or to save time. All technologies have to provide at least one (ideally all three) of these benefits to their users for the overhead of using them. This is true for all tools and technologies from the primitive tools used to create fire to modern computers used to solve all kinds of problems.

Qualities 🔗

Once a problem is solved by a tool, people expect additional attributes to reduce the overhead of using it. This can be seen with almost every new technology, the early versions of computers had to be programmed with punch cards, and people expected (and got) better methods for programming. In engineering, these attributes which describe the overall property of a system, as opposed to the function of the system, are called Non-functional requirements [NFR].

For this post, I’ll be focusing on the NFRs of a product/system that is relevant to the end-user. Some common examples of NFRs relevant to the user are usability, privacy, accessibility, performance. NFRs are also referred to as qualities, constraints, and many other names. I’ll also be using the term “Quality attributes” which are the attributes of a system to meet the qualities or NFRs.

Qualities of a system often do not solve any new problem on their own. But rather they make the process of using the system to solve the problem cheaper, efficient, simpler, etc. After all, all the programming that is done today could have been theoretically done on punch cards, but the introduction of newer programming languages definitely made programming simpler among other things.

Qualities often have tradeoffs with other qualities. This is often expressed by things such as the classic choose two out of three of CAP theorem, or the trade-off between security and usability. Different user groups may prefer different tradeoffs. This is one of the reasons why multiple products exist for the same job since they provide a different tradeoff preferred by a different group of users.

Qualities and Platform 🔗

If we focus on privacy as a quality of a system, we can see a spectrum of services that make the tradeoff of privacy with other qualities. On the one side, we have services like password managers which offer strong privacy guarantees all the way to services that rely on the information you provide in opaque ways to make money on the other. For most services, privacy is just one of the qualities they provide and it may or may not be core to the job they are doing. But for some services, privacy is a critical part of their offering and the business would cease to exist if they default on that quality. This leads these services to focus extensively on getting privacy and security right at the expense of potentially other jobs and in doing that they end up with a pure platform focused exclusively on privacy and security.

Focusing exclusively on providing qualities almost definitionally leads to building platforms. This is because as mentioned earlier qualities by themselves don’t provide any value; rather they have to be paired with a job to extract the value out of the quality. Having a platform has the usual advantage of being able to build and offer multiple products on top of it with reduced time of development and potentially providing the platform as a service and extracting platform fees as a source of revenue. Platforms also end up providing good abstractions which enable tackling more problems at a higher level. Platforms in this sense are a condensed set of qualities which themselves are a set of common expectations from the users.

1Password is a good example of this. It started as a password manager, but it is now expanding to be an enterprise secret manager platform.

Admittedly this way of creating a new platform is a bit atypical since a platform needs to provide not just one quality but multiple qualities for it to be worthwhile incurring the overhead of building on top of it. It is also true that most products do not focus exclusively on a single quality. But the point I’m trying to drive towards is that qualities of a system are usually platform pieces, which just like a platform need to have a product built with it to be useful.

Products and Societal Expectations of Quality 🔗

Although some technologies make certain quality attributes table stakes, the quality expectations for a product fundamentally originate from user expectations. User expectations in turn are molded by time and their environment. Because of this indirect linking to the time and societal expectations most products get baked with quality attributes of the time they were built in. We usually see this as an absence of certain quality attributes for some of the older technologies we use. Security for example was not a primary quality consideration during the development of email. This has led to a patchwork of things being bolted on top of the email protocol (like DMARC, SPF, etc.) to bring it into the current era where security is generally a larger focus.

Unlike protocols that get interwoven with the ecosystem, individual products usually aren’t so lucky. Changing quality expectations are usually hard for products to navigate for a variety of reasons. For one quality is an architecturally significant requirement. This makes it extremely hard to change the product without entirely changing the underlying architecture. And since over time products functionality constraints business models, business models also end up having a transitive coupling with quality expectations of the time.

This means a significant change in quality expectation often leads to disruptions in the business model of products. These business models might still solve the actual job they were meant to just fine but fail to meet the newer quality expectations. We can see this today more starkly in legacy social media companies like Facebook. Facebook continues to solve the problem of connecting people and it in fact might do it better than it did before. But it was built with the expectation users would provide it with their information and share their interests which it could then use to monetize them via ads. The growing privacy moment today is changing that expectation and making privacy a more important quality expectation. Apple which started with privacy as a quality attribute across its product happily fans the flames via marketing and messaging just like any good competition. This messaging often strengthens and ends up as input for the quality expectations of society. Although this is not the only reason for the backlash of Facebook among people, it certainly is one of them.

Quality expectations for a group usually change over a long period. But it forms a major force in disrupting existing business models which work perfectly fine until they don’t.